
%run "/code/source/config/notebook_settings.py"


pd.set_option('display.max_colwidth', None)


from source.library.text_analysis import count_tokens, tf_idf, get_context_from_keyword, \
    count_keywords, count_keywords_by, impurity


with Timer("Loading Data"):
    path = 'artifacts/data/processed/reddit.pkl'
    df = pd.read_pickle(path)

2023-02-17 07:41:11 - INFO     | Timer Started: Loading Data
2023-02-17 07:41:11 - INFO     | Timer Finished: (0.24 seconds)


df.head(1)


hlp.pandas.numeric_summary(df)


hlp.pandas.non_numeric_summary(df)


df['post'].iloc[0][0:1000]

"Hello, <lb><lb>I have my J32A3 egine open, ready to put new rings and all. I would like to know if I can swap the J35Z2 crank and rods from an Accord Sedan v6 3.5. I would like to keep my pistons, just change the rods and crank to have a little more displacement. This is a doable option? there is something I need to know first? This is for my Acura TL 3G 2004 Manual Transmission. I really would like to know if it's possible and if possible what differences are between then J35Z2 and let's say the classic J35A3.<lb><lb>Thanks in advance."


'|'.join(df['partial_lemmas'].iloc[0])[0:1000]

'hello|j32a3|egine|open|ready|new|ring|like|know|swap|j35z2|crank|rod|accord|sedan|v6|_number_|like|piston|change|rod|crank|little|displacement|doable|option|need|know|acura|tl|3|g|_number_|manual|transmission|like|know|possible|possible|difference|j35z2|let|classic|j35a3|thank|advance'


'|'.join(df['bi_grams'].iloc[0])[0:1000]

'j32a3 egine|egine open|new ring|j35z2 crank|accord sedan|sedan v6|v6 _number_|doable option|acura tl|tl 3|3 g|g _number_|_number_ manual|manual transmission|classic j35a3'


'|'.join(df['noun_phrases'].iloc[0])[0:1000]

'egine open|new ring|j35z2 crank|doable option|classic j35a3'


ax = df['impurity'].plot(kind='box', vert=False, figsize=(10, 1))
ax.set_title("Distribution of Post Impurity")
ax.set_xlabel("Impurity")
ax.set_yticklabels([])
ax;


df[['impurity', 'post', 'post_clean']].sort_values('impurity', ascending=False).head()


df['language'].value_counts(ascending=False)

English    4996
Name: language, dtype: int64


df['subreddit'].value_counts(ascending=False)

Lexus                 266
Hyundai               263
Trucks                262
Honda                 261
MPSelectMiniOwners    260
mercedes_benz         259
mazda3                257
Harley                255
volt                  252
Volkswagen            252
Audi                  252
teslamotors           250
Volvo                 249
Mustang               248
BMW                   239
saab                  239
4Runner               238
Porsche               236
subaru                233
Wrangler              229
Name: subreddit, dtype: int64


remove_tokens = {'_number_', 'car'}
count_tokens(df['partial_lemmas'], remove_tokens=remove_tokens).head(10)


ax = df['post_length'].plot(kind='box', vert=False, figsize=(10, 1))
ax.set_title("Distribution of Post Length")
ax.set_xlabel("# of Characters")
ax.set_yticklabels([])
ax;


ax = df['post_length'].plot(kind='hist', bins=60, figsize=(10, 2));
ax.set_title("Distribution of Post Length")
ax.set_xlabel("# of Characters")
ax;


import seaborn as sns
sns.displot(df['post_length'], bins=60, kde=True, height=3, aspect=3);


where = df['subreddit'].isin([
    'Lexus', 
    'mercedes_benz',
    'Audi',
    'Volvo',
    'BMW',
])
g = sns.catplot(data=df[where], x="subreddit", y="post_length", kind='box')
g.fig.set_size_inches(6, 3)
g.fig.set_dpi(100)
g = sns.catplot(data=df[where], x="subreddit", y="post_length", kind='violin')
g.fig.set_size_inches(6, 3)
g.fig.set_dpi(100)


counts_df = count_tokens(df['partial_lemmas'], remove_tokens=remove_tokens)


def plot_wordcloud(frequency_dict):
    wc = wordcloud.WordCloud(background_color='white',
        #colormap='RdYlGn',
        colormap='tab20b',
        width=round(hlp.plot.STANDARD_WIDTH*100),
        height=round(hlp.plot.STANDARD_HEIGHT*100),
        max_words = 200, max_font_size=150,
        random_state=42
    )
    wc.generate_from_frequencies(frequency_dict)

    fig, ax = plt.subplots(figsize=(hlp.plot.STANDARD_WIDTH, hlp.plot.STANDARD_HEIGHT))
    ax.imshow(wc, interpolation='bilinear')
    #plt.title("XXX")
    plt.axis('off')


plot_wordcloud(counts_df.to_dict()['frequency']);


tf_idf_lemmas = tf_idf(
    df=df,
    tokens_column='partial_lemmas',
    segment_columns = None,
    min_frequency_corpus=20,
    min_frequency_document=20,
    remove_tokens=remove_tokens,
)
tf_idf_lemmas.head()


remove_tokens_bi_grams = {'_number_ year', '_number_ _number_', 'hey guy'}
tf_idf_bi_grams = tf_idf(
    df=df,
    tokens_column='bi_grams',
    segment_columns = None,
    min_frequency_corpus=20,
    min_frequency_document=20,
    remove_tokens=remove_tokens_bi_grams,
)
tf_idf_bi_grams.head()


tf_idf_nouns = tf_idf(
    df=df,
    tokens_column='nouns',
    segment_columns = None,
    min_frequency_corpus=20,
    min_frequency_document=20,
    remove_tokens=remove_tokens,
)
tf_idf_nouns.head()


tf_idf_noun_phrases = tf_idf(
    df=df,
    tokens_column='noun_phrases',
    segment_columns = None,
    min_frequency_corpus=20,
    min_frequency_document=20,
    remove_tokens=remove_tokens_bi_grams,
)
tf_idf_noun_phrases.head()


ax = tf_idf_lemmas.head(30)[['tf-idf']].plot(kind='barh', width=0.99)
ax.set_title("TF-IDF of Uni-Grams")
ax.set_xlabel("TF-IDF")
ax.invert_yaxis();


ax = tf_idf_bi_grams.head(30)[['tf-idf']].plot(kind='barh', width=0.99)
ax.set_title("TF-IDF of Bi-Grams")
ax.set_xlabel("TF-IDF")
ax.invert_yaxis();


ax = tf_idf_nouns.head(30)[['tf-idf']].plot(kind='barh', width=0.99)
ax.set_title("TF-IDF of Bi-Grams")
ax.set_xlabel("TF-IDF")
ax.invert_yaxis();


ax = tf_idf_noun_phrases.head(30)[['tf-idf']].plot(kind='barh', width=0.99)
ax.set_title("TF-IDF of Bi-Grams")
ax.set_xlabel("TF-IDF")
ax.invert_yaxis();


plot_wordcloud(tf_idf_lemmas.to_dict()['tf-idf']);


plot_wordcloud(tf_idf_bi_grams.to_dict()['tf-idf']);


remove_tokens_subreddit = set(df.subreddit.str.lower().unique())
remove_tokens_subreddit

{'4runner',
 'audi',
 'bmw',
 'harley',
 'honda',
 'hyundai',
 'lexus',
 'mazda3',
 'mercedes_benz',
 'mpselectminiowners',
 'mustang',
 'porsche',
 'saab',
 'subaru',
 'teslamotors',
 'trucks',
 'volkswagen',
 'volt',
 'volvo',
 'wrangler'}


tf_idf_lemmas_per_sub = tf_idf(
    df=df,
    tokens_column='partial_lemmas',
    segment_columns = 'subreddit',
    min_frequency_corpus=10,
    min_frequency_document=10,
    remove_tokens=remove_tokens | remove_tokens_subreddit 
)
tf_idf_lemmas_per_sub.head(5)


tf_idf_bigrams_per_sub = tf_idf(
    df=df,
    tokens_column='bi_grams',
    segment_columns = 'subreddit',
    min_frequency_corpus=10,
    min_frequency_document=10,
    remove_tokens=remove_tokens_bi_grams
)
tf_idf_bigrams_per_sub.head(5)


tf_idf_nouns_per_sub = tf_idf(
    df=df,
    tokens_column='nouns',
    segment_columns = 'subreddit',
    min_frequency_corpus=10,
    min_frequency_document=10,
    remove_tokens=remove_tokens | remove_tokens_subreddit
)
tf_idf_nouns_per_sub.head(5)


tf_idf_nounphrases_per_sub = tf_idf(
    df=df,
    tokens_column='noun_phrases',
    segment_columns = 'subreddit',
    min_frequency_corpus=10,
    min_frequency_document=10,
    remove_tokens=remove_tokens_bi_grams
)
tf_idf_nounphrases_per_sub.head(5)


tokens_to_show = tf_idf_lemmas_per_sub.query("subreddit in ['Lexus', 'Volvo']").reset_index()
tokens_to_show.head()


px.bar(
    tokens_to_show.groupby(['subreddit']).head(20).sort_values('tf-idf', ascending=True),
    x='tf-idf',
    y='token',
    color='subreddit',
    barmode='group',
    title="Top 20 Lemmas for Volvo & Lexus"
)


tokens_to_show = tf_idf_bigrams_per_sub.query("subreddit in ['Lexus', 'Volvo']").reset_index()
tokens_to_show.head()


px.bar(
    tokens_to_show.groupby(['subreddit']).head(20).sort_values('tf-idf', ascending=True),
    x='tf-idf',
    y='token',
    color='subreddit',
    barmode='group',
    title="Top 20 Bi-Grams for Volvo & Lexus"
)


get_context_from_keyword(df.query("subreddit == 'Lexus'")['post'], keyword='think')

1010    ethod about 2 years ago. I used to  |think|  it happened more with a light thro
1010    replace?<lb><lb>Question 2) Do you  |think|  a passageway in the transmission p
3870     cheap and easy fix or do you guys  |think|  an insurance claim will end up nee
2657    s subreddit regularly and couldn't  |think|  of a better group of people to ask
1061     what were your reasons? Overall I  |think|  both the cars are pretty evenly ma
3718     and my spouse.<lb><lb>Do you guys  |think|  the tire/wheel protection was wort
1190    e on the 2005-2006 year. I am also  |think| ing about the GS. since I'm assumin
4399    e. They smell nice, which makes me  |think|  that it's an air freshener or some
4399    ere would I even put them? I don't  |think|  it's candy. <lb><lb>https://i.imgu
1283    d it's fine, my question is do you  |think|  the battery could have been draine
dtype: object


get_context_from_keyword(df.query("subreddit == 'Volvo'")['post'], keyword='think')

3411     the GLE model.  Anyways I've been  |think| ing about purchasing this Volvo 740
2208    est in getting it fixed. I started  |think| ing if my C70 and his V70 are simil
2544    iously considering a Volvo V60. We  |think|  we can afford a 4 year old car wit
1202    ination of settings in menus I can  |think|  of, balance, faders, equalisers, 3
1543    n about 35,000 miles, but we don't  |think|  it's driven more than 60,000 or so
1314    d reviews, so I don't know what to  |think| . In fact, I don't even know what t
1194    wrench even came out, and I didn't  |think|  they were that tight. This can't b
4612     valve seal issue. What do you all  |think| ? Is the cost of repair worth this 
4099    mes on.  I try to stay positive by  |think| ing maybe my mechanic didn't top me
2134    etween gears while moving. I don't  |think|  the clutch is fully disengaging fr
dtype: object


tokens_to_show = tf_idf_lemmas_per_sub.query("subreddit == 'Lexus'").reset_index()
#tokens_to_show = tokens_to_show[~tokens_to_show.token.isin(stop_words)]
tokens_to_show = tokens_to_show[['token', 'tf-idf']].set_index('token')
tokens_to_show = tokens_to_show.to_dict()['tf-idf']
plot_wordcloud(tokens_to_show);


tokens_to_show = tf_idf_lemmas_per_sub.query("subreddit == 'Volvo'").reset_index()
#tokens_to_show = tokens_to_show[~tokens_to_show.token.isin(stop_words)]
tokens_to_show = tokens_to_show[['token', 'tf-idf']].set_index('token')
tokens_to_show = tokens_to_show.to_dict()['tf-idf']
plot_wordcloud(tokens_to_show);


contexts = get_context_from_keyword(
    documents=df['post'],
    window_width=50,
    keyword='replac',
    num_samples = 20,
    random_seed=42
)
for x in contexts:
    print(x)

rger.<lb>* My Elantra still runs great and I just  |replac| ed the tires.<lb>* The financials are all there, b
 for a while, but now the time has come for me to  |replac| e my coil springs. I had heard great things about 
I'm  |replac| ing the radio with a Pioneer AVH-X2800BS in my '04
<lb><lb>I sent it back to PA Performance and they  |replac| ed all the internals. While I was waiting on PA Pe
 Also, the battery in the car seems like it needs  |replac| ing as the interior lights flicker and the car has
ere can confirm that before I go digging deep for  |replac| ement panels. Thanks!
and my steering stabilizer is shot.  Do I need to  |replac| e it to keep my tires healthy or not?  I really do
nyone knows the type of them? Or where can i find  |replac| ements? I will add pictures, the small one is from
 is having some problems so that might need to be  |replac| ed. Could that explain the floatiness? <lb><lb>Als
<lb>The coolant reservoir tank seems to have been  |replac| ed, I heard it's a common problem on these.
Anyone else experience this? Dealer quote to  |replac| e it was insane for what I imagine is a relatively
I know the gasket needs to be  |replac| ed and the mechanic i take mine too agreed to let 
I also asked Mazda service dealers for a quote in  |replac| ing this part. The quotes varied wildly with some 
 it not for corrosion around the old part.<lb><lb> |Replac| ing the vehicle speed sensor, or driven-speed gear
lb>My boss has a 2010 Dodge Ram 1500 and wants to  |replac| e his old, corroded bars with some new ones. Thing
 the things that I am seeing is pretty much about  |replac| ing the entire intake manifold. Also seeing a lot 
My fault, I fucked up  |replac| ing the door lock actuator and now my rear passeng
 may bite the bullet and buy a used door panel to  |replac| e it with unless you guys have any suggestions. 
gs, they were original, appear to have aged well.  |Replac| ed the plugs, and switched the coil pack from 1-2 
 every now and then it simply won't start. I have  |replac| ed and checked the battery, but I'm thinking I may

	# of Non-Nulls	% Nulls	# of Zeros	% Zeros	Mean	St Dev.	Coef of Var	Skewness	Kurtosis	Min	10%	25%	50%	75%	90%	Max
impurity	5,000	0.0%	1,023	20.0%	0.0	0.0	1.0	1.9	7.8	0.0	0.0	0.0	0.0	0.0	0.0	0.2
post_length	5,000	0.0%	0	0.0%	679.1	452.5	0.7	2.5	9.2	256.0	307.0	381.0	538.0	812.2	1,217.1	4,174.0
num_tokens	5,000	0.0%	0	0.0%	54.2	36.3	0.7	2.6	9.9	15.0	25.0	31.0	43.0	65.0	96.0	378.0

	# of Non-Nulls	# of Nulls	% Nulls	Most Freq. Value	# of Unique	% Unique
id	5,000	0	0.0%	74qv99	5,000	100.0%
subreddit	5,000	0	0.0%	Lexus	20	0.4%
title	5,000	0	0.0%	Need some advice	4,995	99.9%
post	5,000	0	0.0%	Hello, <lb><lb>I have my J32A3[...]	5,000	100.0%
post_clean	5,000	0	0.0%	Hello, I have my J32A3 egine o[...]	5,000	100.0%
all_lemmas	5,000	0	0.0%	['hello', 'i', 'have', 'my', '[...]	5,000	100.0%
partial_lemmas	5,000	0	0.0%	['hello', 'j32a3', 'egine', 'o[...]	5,000	100.0%
bi_grams	5,000	0	0.0%	['j32a3 egine', 'egine open', [...]	5,000	100.0%
adjs_verbs	5,000	0	0.0%	['open', 'ready', 'new', 'like[...]	5,000	100.0%
nouns	5,000	0	0.0%	['j32a3', 'egine', 'ring', 'j3[...]	5,000	100.0%
noun_phrases	5,000	0	0.0%	[]	4,981	99.6%
entities	5,000	0	0.0%	[]	4,559	91.2%
language	4,996	4	0.1%	English	1	0.0%

	impurity	post	post_clean
4684	0.18	I'm looking to lease an a4 premium plus automatic with the nav package.<lb><lb>Vehicle Price:<tab><tab>$49,150.00<tab> <lb> <tab>AutoNation Savings:<tab>-<tab>$3,867.00<tab> <lb> <tab>AutoNation Price:<tab><tab>$45,283.00<tab> <lb> <tab> <tab> <lb> <tab>Sales Tax (estimate):<tab>+<tab>$2,734.98<tab> <lb> <tab>Title Fee:<tab>+<tab>$100.00<tab> <lb> <tab>Tire/Battery/MVWEA:<tab>+<tab>$4.00<tab> <lb> <tab>Tag/Registration Fees (estimate):<tab>+<tab>$207.00<tab> <lb> <tab>Electronic Filing:<tab>+<tab>$20.00<tab> <lb> <tab>Other:<tab>+<tab>$20.00<tab> <lb> <tab>Documentation Fee:<tab>+<tab>$300.00<tab> <lb> <tab>Balance Due (estimate):<tab><tab>$48,668.98<tab> <tab>No Trade-In<lb><lb>LEASE OPTIONS<lb>Cash Due<tab>36 months <tab>42 months <lb><lb>$2,000 <tab>$723<tab>$690<lb>$4,000 <tab>$663<tab>$639<lb>$6,000 <tab>$603<tab>$587<lb><lb><lb>This is my first lease, do these numbers look good? Should I push back or negotiate on anything?<lb><lb>Thanks!	I'm looking to lease an a4 premium plus automatic with the nav package. Vehicle Price: $ _NUMBER_ AutoNation Savings: $ _NUMBER_ AutoNation Price: $ _NUMBER_ Sales Tax (estimate): $ _NUMBER_ Title Fee: $ _NUMBER_ Tire/Battery/MVWEA: $ _NUMBER_ Tag/Registration Fees (estimate): $ _NUMBER_ Electronic Filing: $ _NUMBER_ Other: $ _NUMBER_ Documentation Fee: $ _NUMBER_ Balance Due (estimate): $ _NUMBER_ No Trade-In LEASE OPTIONS Cash Due _NUMBER_ months _NUMBER_ months $ _NUMBER_ $ _NUMBER_ $ _NUMBER_ $ _NUMBER_ $ _NUMBER_ $ _NUMBER_ $ _NUMBER_ $ _NUMBER_ $ _NUMBER_ This is my first lease, do these numbers look good? Should I push back or negotiate on anything? Thanks!
1287	0.17	Bulbs Needed:<lb><lb><lb>194 LED BULB x8<lb><lb>4- DOORS<lb><lb>2- MAP LIGHTS<lb><lb>2- VANITY<lb><lb><lb>3022 LED BULB x3<lb><lb>2- CARGO DOOR<lb><lb>1- DOME LIGHT<lb><lb><lb>BULBS USED:<lb><lb>[194 LED BULBS](https://goo.gl/Jfu2Dx)<lb><lb>[3022 LED BULBS](https://goo.gl/fPgk6n)<lb><lb>[Trim Tools](https://goo.gl/hjxw8Z)<lb><lb>Parts list courtesy of [The Blue TRD](https://www.youtube.com/watch?v=CBJxfWdbEfo&t=28s) from his You Tube Channel.<lb><lb>Just passing along the helpful info.	Bulbs Needed: _NUMBER_ LED BULB x8 _NUMBER_ - DOORS _NUMBER_ - MAP LIGHTS _NUMBER_ - VANITY _NUMBER_ LED BULB x3 _NUMBER_ - CARGO DOOR _NUMBER_ - DOME LIGHT BULBS USED: _NUMBER_ LED BULBS _NUMBER_ LED BULBS Trim Tools Parts list courtesy of The Blue TRD from his You Tube Channel. Just passing along the helpful info.
142	0.15	Breakdown below:<lb><lb>Elantra GT<lb><lb>2.0L 4-cylinder<lb><lb>6-speed Manual Transmission<lb><lb>$19,350<lb><lb>Elantra GT<lb><lb>2.0L 4-cylinder<lb><lb>6-speed Automatic Transmission w/ SHIFTRONIC®<lb><lb>$20,350<lb><lb>Elantra GT Sport<lb><lb>1.6L Turbo GDI 4-cylinder<lb><lb>6-speed Manual Transmission<lb><lb>$23,250<lb><lb>Elantra GT Sport<lb><lb>1.6L Turbo GDI 4-cylinder<lb><lb>7-speed EcoShift® Dual Clutch Transmission w/ SHIFTRONIC®<lb><lb>$24,350	Breakdown below: Elantra GT _NUMBER_ .0L _NUMBER_ -cylinder _NUMBER_ -speed Manual Transmission $ _NUMBER_ Elantra GT _NUMBER_ .0L _NUMBER_ -cylinder _NUMBER_ -speed Automatic Transmission w/ SHIFTRONIC® $ _NUMBER_ Elantra GT Sport _NUMBER_ .6L Turbo GDI _NUMBER_ -cylinder _NUMBER_ -speed Manual Transmission $ _NUMBER_ Elantra GT Sport _NUMBER_ .6L Turbo GDI _NUMBER_ -cylinder _NUMBER_ -speed EcoShift® Dual Clutch Transmission w/ SHIFTRONIC® $ _NUMBER_
3174	0.13	E-price:<lb>$20,863.00<lb>Freight:<lb>$900.00<lb>Processing Fee:<lb>$299.00<lb>Total before tax and tag fees:<lb>$22,062.00<lb>7% State SALES TAX: $ 1,544.34<lb>2 YEAR TAG FEES: $187.00<lb>TITLE: $100.00<lb>REGISTRATION: $20.00<lb>LIEN: $20.00<lb>INSPECTION: $25.00 <lb>State TEMP TAG: $20.00<lb>State TIRE FEE: $4.00<lb>TOTAL OUT THE DOOR YOU REQUESTED: $ 23,982.34<lb>	E-price: $ _NUMBER_ Freight: $ _NUMBER_ Processing Fee: $ _NUMBER_ Total before tax and tag fees: $ _NUMBER_ _NUMBER_ % State SALES TAX: $ _NUMBER_ _NUMBER_ YEAR TAG FEES: $ _NUMBER_ TITLE: $ _NUMBER_ REGISTRATION: $ _NUMBER_ LIEN: $ _NUMBER_ INSPECTION: $ _NUMBER_ State TEMP TAG: $ _NUMBER_ State TIRE FEE: $ _NUMBER_ TOTAL OUT THE DOOR YOU REQUESTED: $ _NUMBER_
3678	0.12	The lease on my 2014 C250 is ending and I have the option to buy it for $20k.<lb>So I decided to see what else is out there for about the same amount of money and so far this are my options:<lb> <lb> <lb>Model: C250<lb>Year: 2014<lb>Miles: 7k<lb>Price: $20k<lb> <lb>Model: E350<lb>Year: 2010<lb>Miles: 59k<lb>Price: $19k<lb> <lb>Model: ML350<lb>Year: 2006<lb>Miles: 42k<lb>Price: $13k<lb> <lb>Thoughts?	The lease on my _NUMBER_ C250 is ending and I have the option to buy it for $20k. So I decided to see what else is out there for about the same amount of money and so far this are my options: Model: C250 Year: _NUMBER_ Miles: 7k Price: $20k Model: E350 Year: _NUMBER_ Miles: 59k Price: $19k Model: ML350 Year: _NUMBER_ Miles: 42k Price: $13k Thoughts?

	frequency	tf-idf
token
look	2776	3043.65
drive	1880	2867.85
like	2355	2830.75
new	1738	2577.63
mile	1406	2510.05

	frequency	tf-idf
token
$ _number_	1089	2366.46
_number_ mile	553	1417.20
look like	229	749.94
_number_ month	194	676.54
test drive	188	670.66

Citation¶

Setup¶

Exploratory Data Analysis¶

Dataset Summary¶

Numeric Summary¶

Non-Numeric¶

Examples¶

Explore Non-Text Columns¶

Impurity¶

Explore Text Column|¶

Top Words Used¶

Distribution of Text Length¶

Word Frequency¶

TF-IDF¶

By Subreddit¶

Lexus¶

Volvo¶

Keywords in Context¶

	frequency	tf-idf
token
mile	1404	2507.99
issue	1142	2240.32
year	1227	2226.76
time	1289	2216.42
engine	944	2100.66

	frequency	tf-idf
token
oil change	132	514.63
new car	130	502.27
test drive	110	442.43
engine light	101	439.81
first time	109	433.04

		frequency	tf-idf
subreddit	token
4Runner	gen	74	283.40
	sr5	50	241.65
	lift	57	215.97
	rear	61	171.54
	look	148	162.27

	subreddit	token	frequency	tf-idf
0	Lexus	is350	37	198.29
1	Lexus	look	166	182.01
2	Lexus	mile	101	180.31
3	Lexus	gs	30	160.77
4	Lexus	drive	103	157.12

	subreddit	token	frequency	tf-idf
0	Lexus	$ _number_	54	117.35
1	Lexus	_number_ lexus	19	105.00
2	Lexus	f sport	16	94.55
3	Lexus	_number_ is250	13	82.09
4	Lexus	es _number_	13	80.85

		frequency	tf-idf
subreddit	token
4Runner	sway bar	12	65.78
4Runner	check engine	10	44.13
Harley	new bike	14	84.73
	first bike	11	68.41
	spark plug	10	46.85